As NCI researchers, many of us routinely work with large sequencing datasets, high-resolution imaging files, molecular simulations, or machine learning models. These analyses quickly exceed the capacity of a typical desktop. That’s where high-performance computing (HPC) comes in.
What is HPC?
High-performance computing (HPC) refers to the use of interconnected computers (clusters) that work in parallel to process large or complex computational jobs efficiently. Instead of running analyses on a single machine, HPC distributes work across many processors, dramatically reducing runtime and enabling large-scale analysis.
Within the NIH and NCI ecosystem, two major HPC environments are available to CCR researchers:
NIH HPC: Biowulf & Helix
The NIH HPC System supports intramural researchers primarily through Biowulf and Helix.
Biowulf
Biowulf is NIH’s flagship Linux-based HPC cluster designed for large-scale, high-throughput scientific computing. It provides:
- Thousands of CPU cores
- High-memory nodes
- GPU nodes for AI and deep learning
- Petabyte-scale storage
- Support for thousands of simultaneous jobs
Hundreds of scientific software packages are pre-installed and managed through the module system, including tools for genomics, structural biology, image analysis, statistics, and machine learning.
Biowulf also provides HPC Open OnDemand, a web-based portal that allows users to:
- Launch interactive desktop sessions
- Run Jupyter notebooks, RStudio, and other interactive tools
- Access command-line tools without SSH
This interface is especially helpful for users who prefer GUI-based workflows or are new to HPC.
Helix
Helix is the interactive companion system to Biowulf and is commonly used for data transfer and file management.
Documentation & Training
One of Biowulf’s major strengths is its extensive documentation and training resources. The NIH HPC website includes:
- Detailed user guides https://hpc.nih.gov/docs/user_guides.html
- Example job submission scripts
- Software-specific documentation https://hpc.nih.gov/apps/
- Training workshops and user consultations
If you are new to HPC or troubleshooting workflows, the Biowulf documentation is particularly comprehensive and user-friendly.
Getting an NIH HPC Account
NIH employees, contractors, and affiliates listed in the NIH Enterprise Directory (NED) are eligible.
To request an account:
- Submit an account request form via the NIH HPC website
- Obtain approval from your Principal Investigator
- Renew annually
There is a monthly maintenance fee ($40) associated with NIH HPC accounts.
Frederick Research Computing Environment (FRCE)
🔗 https://ncifrederick.cancer.gov/staff/FRCE
The Frederick Research Computing Environment (FRCE) is a Linux-based HPC cluster that supports NCI research, particularly at the Frederick campus, but is available to NCI researchers more broadly.
What FRCE Offers
FRCE provides:
- Thousands of CPU cores
- GPU resources
- Slurm job scheduling
- Software for genomics, imaging, and computational biology.
- Open OnDemand web portal for browser-based interactive access
FRCE supports both batch and interactive computing workflows and is well integrated with Frederick-based research infrastructure.
Note: FRCE has less computational resources than Biowulf.
Getting a FRCE Account
NCI researchers are eligible for a no-cost FRCE account.
To request access:
- Visit the FRCE website and follow the account request instructions https://ncifrederick.cancer.gov/staff/FRCE/Support/AccessRequest.
- Contact the FRCE support team or submit a ServiceNow ticket for assistance https://ncifrederick.cancer.gov/staff/FRCE
FRCE Training and Documentation
Check out the documentation for use guides and other information to get started with FRCE. https://ncifrederick.cancer.gov/staff/FRCE/Documentation.
Choosing the Right Resource
Both Biowulf and FRCE:
- Use the Slurm job scheduler
- Provide Open OnDemand web access
- Support CPU- and GPU-based workflows
- Provide access to DRAGEN hardware-accelerated genomics pipelines
- Offer user support
In general:
- Choose Biowulf for very large-scale workloads, extensive documentation, and NIH-wide infrastructure.
- Choose FRCE for NCI-focused workflows, Frederick integration, and no-cost access.
As a bioinformatics trainer, my advice is simple: if your analysis is too big for your laptop, HPC is likely the right next step. Both environments are robust, well-supported, and ready to accelerate your research.
If your analysis needs exceed local HPC capacity or would benefit from elastic, cloud-native infrastructure, NIH STRIDES offers supported access to commercial cloud platforms for scalable research computing.
– BTEP Team